Closelook Daily Pulse

Temperature 59 Risk-On as AW25 Surges 5.3% in Single Session

Thu, 16 Apr 2026 00:00:00 GMT

The Closelook Temperature sits at 59 🟡, labelled Risk-on rally, indicating broad appetite for risk assets without yet reaching the elevated heat that historically precedes exhaustion.

The session's dominant move came from AW25, the agentic-winners basket, which gained 5.3% on the day. That single-day print stands out against a muted one-week reading of -1.5%, suggesting the basket had been under distribution before Wednesday's sharp reversal. The move warrants attention: agentic-native names can reprice abruptly when sentiment shifts, and a 5% gap without a directional week behind it often signals repositioning rather than a clean trend resumption.

Elsewhere the picture was mixed. The Rubin 100 slipped 0.75% on the day but retains a strong 6.7% one-week gain and an exceptional 22.5% one-month move, reflecting the sustained build-out cycle in AI infrastructure. HALO 100 added 0.56% and is up 2.4% on the week, though the one-month reading of -1.7% shows that large-cap defensive tech continues to lag the infrastructure and agentic cohorts. Euro-AI 50 was essentially flat on the day, up 2.8% on the week and 6.8% on the month — the steadiest performer across the four indices.

The Pattern Scanner recorded 25 active signals, all clustering around support-confluence setups. The top names — MCD, XOM, UL, PLTR, SNY — span defensives, energy, and a single high-beta name in Palantir. The dominance of support-confluence patterns rather than breakout or momentum patterns is consistent with a market that is recovering from a drawdown phase and testing whether prior support levels hold as new floors.

Cointegration stress is elevated: zero active pairs and six recent breaks. That level of pair dislocation points to cross-asset correlations still in flux, which can create false signals in relative-value frameworks. We treat this as a reason for extra caution on mean-reversion exposures until pair structures begin to re-establish.

Watch whether AW25 can consolidate above Wednesday's open over the next two sessions, and whether the Rubin 100's one-month momentum holds as earnings season accelerates into the back half of April.

Introducing the Agentic Winners 25

Tue, 31 Mar 2026 00:00:00 GMT

We have built the large strategic multi-year indices — Rubin Build-Out 100, Euro-AI Sovereign 50, HALO 50. Now we complement them with smaller tactical indices suited for trading horizons of six months. The Closelook Agentic Winners 25 is the first.

25 names across 7 sectors: MAG7 Gateway (MSFT, GOOGL, AAPL), Platform (SAP, NOW, CRM, WDAY), Toll Booth (NET, DDOG, CRWD, TWLO, OKTA), Neocloud/Edge (DOCN, FSLY, QCOM), Data/Search (SNOW, MDB, ESTC), Vertical (VEEV, INTU, PTC, CSGP, IOT), Commerce + Orchestration (SHOP, PLTR).

The thesis: at some point in H1 2026, all the negatives get priced into enterprise software. The stocks that form a base will be the first-generation agentic winners — the platforms agents cannot operate without.

The current chart picture tells an early story. PLTR shows relative strength — the strongest AI software stock during the upmove, holding its ascending long-term trendline while the sector sells off. Expected to lead the next cycle. MSFT is at a critical multi-year support zone near $356, showing no sign of a floor yet. IGV, the broad software ETF, is approaching a potential double bottom near $73–76 — the same level we flagged in February.

The practical test we are watching: when NDX drops 2%, does this basket drop 3% (fear dominates, beta-amplified selling) or 1.5% (strategic accumulation absorbing selling pressure)? As long as the group underperforms on down days, the AI disruption fear narrative is in control. The regime shifts when the group starts dropping less than the index on selloff days and recovering faster on bounce days.

The combination signal: Rubin Build-Out pauses or consolidates as the infrastructure capex cycle matures, while software names begin showing relative outperformance — double bottoms forming on long-term charts, support zones holding where they previously broke. Price action shows before news.

We will publish 7 sector sub-index charts and beta-adjusted relative strength data as the AW25 tracking goes live.

Jensen Confirms the Inference Inflection

Tue, 31 Mar 2026 00:00:00 GMT

NVIDIA invested $2 billion in Marvell today, built around NVLink Fusion — a rack-scale platform enabling semi-custom AI inference infrastructure. Marvell supplies custom XPUs and scale-up networking; NVIDIA contributes Vera CPU, ConnectX, BlueField, NVLink interconnect, Spectrum-X switches.

Jensen's framing is precise: 'The inference inflection has arrived. Token generation demand is surging, and the world is racing to build AI factories.' Not training. Inference. Token generation. That's agents consuming compute in production.

This confirms the structural thesis behind the Closelook Lab tools we launched today:

1. The Agentic Winners 25 tracks the demand side — the 25 software companies that agents cannot operate without. If inference demand is surging, every AW25 sector benefits: Toll Booth (NET, DDOG, CRWD) collects rent per API call. Platform (NOW, CRM, SAP) becomes the operating layer. Commerce + Orchestration (PLTR, SHOP) coordinates the workflows.

2. The Money Temperature dashboard reads 38 — cool regime. SPY and QQQ both at 36. But GLD at 62 and UUP at 58. Gold hot, dollar strong, equities cold. The market hasn't priced the inference inflection yet. When it does, the rotation signal will show in AW25 sector temperatures before it shows in headlines.

3. All 6 cointegration pairs are BREAKING — cascade status CRITICAL. The market is repricing structural relationships. SPY/GLD (p=0.99), TLT/SPY, VEU/SPY — the assumptions behind 60/40, equity-bond correlation, US vs. non-US equilibrium are all dissolved in the current 60-day window.

The Marvell deal adds a hardware lock-in dimension. Custom XPUs that only work within the NVIDIA NVLink ecosystem create switching costs at the silicon level. This is the NVIDIA Revenue Multiplier thesis extending from training into inference.

MRVL +11% on the news. The stock is a Rubin Build-Out 100 constituent (Layer 3 — Networking). Silicon photonics collaboration addresses the interconnect bottleneck that Rubin identifies as the next constraint after packaging and memory.

What we're watching: does today's AW25 basket temperature rise from 38 on the next update? If inference inflection is real and being priced, S3 Toll Booth and S7 Commerce should heat up first.

The AI Optimization Stack Reshapes SaaS

Mon, 30 Mar 2026 00:00:00 GMT

Most investors are still underwriting agent adoption based on today's inference costs. That misses the cost deflation happening at the software layer, not just the model layer.

TurboQuant is one example. But it's just one layer of a broader optimization stack arriving simultaneously: speculative decoding, routing, distillation, sparsity and quantization, context caching. Stacked together, these are multiplicative. A workflow that costs $3–5 today will fall to cents.

The consensus bear case for SaaS has been straightforward: agents disintermediate seat-based software. Lower inference costs flip the question. From "which software gets replaced" to "which platforms are best positioned to deploy agents at scale."

Agents need systems of record, workflow integration, permissions, compliance, security, and proprietary enterprise data. That creates advantages for incumbent platforms with deep workflow embedding, trusted distribution, and data gravity.

The companies most at risk are narrow point solutions replicable via API call. The companies with structural advantages operate within mission-critical workflows and can serve as the operating layer for enterprise agents.

This setup resembles memory stocks before the AI re-rating. What looked like a commodity became strategic. Some of the most interesting AI winners may be enterprise software incumbents the market is still treating as casualties.

The Agentic Economy Is Forming

Sun, 29 Mar 2026 00:00:00 GMT

AI agents are crossing the viability threshold. Not because models got smarter — though they did — but because a wave of software optimizations is about to cut inference costs by 10–20x. A $3–5 AI contract review becomes $0.30. At that price, you run it on everything.

The first viable agents won't replace cheap repetitive tasks. They'll replace expensive human judgment — $500/hour legal reviews, M&A due diligence, clinical workflows. Only high-value tasks absorb messy inference costs today.

The Rubin Build-Out 100 tracks the supply side of AI. Now the demand side is forming. And it's bigger.

We are designing the Closelook Agentic Ecosystem Index: seven layers, 24 sectors, 130–150 companies. Silicon & Architecture, Inference Infrastructure, Toll Booth Infrastructure, Foundation Models & Orchestration, Horizontal Agent Applications, Vertical Agent Applications, and Backdoor Exposure.

What makes this different: a viability timeline overlay. Each sector tagged Phase 1 (viable now), Phase 2 (2027–2029), Phase 3 (commodity inference, 2029+). Plus a two-tier structure: ~90 public companies with price tracking, plus a ~50-company private pipeline.

The Rubin tracks what enables inference. The Agentic Index tracks what inference enables. Together they map the full AI economy.

Europe's AI Story Is in Infrastructure

Fri, 27 Mar 2026 00:00:00 GMT

Europe has an AI equity story. It's not where most people are looking.

The Euro-AI Sovereign 50 tracks 50 European-listed companies across the full AI value chain. No US names. No ADRs. Pure European exposure. Six sectors, each constituent scored with a Relevance Index measuring how central AI is to revenue and strategic positioning.

What the index shows: massive annual performance divergence. Semi Equipment & Materials: +61.7%. Power, Grid & Cooling: +45.7%. Tier 1 Sub-Index: +42.5%. Enterprise AI & Data: −33.9%.

The infrastructure layer is surging — ASML, BE Semiconductor, SÜSS MicroTec, Siemens Energy, Eaton, Schneider Electric. Picks and shovels, European edition.

The enterprise software and data layer is under severe pressure. SAP, UiPath, Capgemini, Dassault Systèmes, Wolters Kluwer — all deeply negative on a 3-month basis. For these companies, AI is increasingly a cost pressure and margin risk rather than a revenue catalyst.

This is the AI Barbell thesis playing out in European equities: infrastructure beneficiaries on one side, a squeezed middle on the other.

One thing to watch: the Enterprise AI & Data sub-index may be forming a double bottom. Too early to confirm, but the chart structure deserves attention.

TurboQuant Changes the Math

Thu, 26 Mar 2026 00:00:00 GMT

Google just compressed AI's working memory by 6x. The market sold memory stocks. That logic is wrong.

TurboQuant is a training-free compression algorithm that quantizes LLM key-value caches to 3 bits with zero accuracy loss. 6x memory reduction. Up to 8x speedup on H100s. No retraining required. Works on any existing model. Drop it in, compress, deploy.

TurboQuant doesn't reduce the need for memory. It makes memory useful enough to unlock an entire category of applications that didn't clear the cost bar before. Same GPU, 6x more concurrent agents. Same phone, dramatically longer context windows.

What happens when inference becomes 6x cheaper overnight? They run 6x more agents.

This is the DeepSeek playbook, one layer up the stack. DeepSeek was about training efficiency. TurboQuant is about inference efficiency — and inference is where the bigger money will be.

Three unlocks: agents become viable on consumer devices (30B model on M5 MacBook). Hyperscaler economics flip from demo to production-viable. And agentic commerce becomes a real business — Google's UCP protocol combined with affordable inference creates consumer-scale agent deployment.

TurboQuant is software. Software runs on silicon. The inference silicon landscape is different from the training landscape — and that distinction matters for portfolio positioning.

The AI Trade Has a Visibility Problem

Wed, 25 Mar 2026 00:00:00 GMT

Everyone watches NVIDIA. Few watch what NVIDIA cannot route around.

The Rubin Build-Out 100 tracks 100 companies across 18 sectors forming the physical infrastructure constraint of the AI buildout. Not chip designers. Not hyperscalers. The companies that deposit, etch, package, interconnect, cool, and power the silicon.

Returns since base (1,000 = October 3, 2022): Rubin EW 4,773 (+377%). Memory 13,239 (+1,224%). Substrates 3,636 (+264%). Advanced Packaging 5,680 (+468%).

YTD leaders: Memory +61.15%. Advanced Materials +46.64%. Packaging +43.55%. High-Speed Interconnects +36.83%. Laggards: EDA & Chip IP (+1.13%), Architects/Chip Design (−7.00%). The design layer is flat while the physical layer rips.

Sector spotlight — S14, High-Speed Interconnects (10 stocks, +36.83% YTD): optical/photonics names leading — AAOI +227%, Furukawa Electric +193%, Lumentum +116%. Electrical interconnect lagging — Amphenol +1%, Astera Labs −15%, Credo −24%.

That intra-sector divergence tells you where the next bottleneck sits: photonic. 800G and 1.6T optical links at data center scale. This is what a functional index surfaces that a cap-weighted ETF never will.

18 sectors. 100 companies. 31 indices. Updated daily after market close.

Most Investors Are Wrong About Gold

Tue, 24 Mar 2026 00:00:00 GMT

Gold is not a reliable inflation hedge. The data is clear.

Since 1971, the World Gold Council reports that only 16% of gold price variation can be explained by changes in US CPI inflation. CFA Institute's 2024 analysis found that changes in headline PCE inflation were not meaningfully correlated with gold on average. The rolling 36-month inflation beta flips sign with confidence intervals including zero.

What gold actually is: a hedge against monetary disorder and real-rate compression, not against inflation itself. No significant cointegration between gold and US CPI from 1971 to 2020, but strong cointegration with the US M2 money supply.

Gold is a hybrid macro asset — highly sensitive to liquidity, money supply, real rates, dollar conditions, and systemic stress. In easy-money regimes, it can behave more like the Nasdaq than the CPI.

The demand structure matters. Total above-ground gold stock at end of 2025: approximately 219,891 tonnes. India and China dominate marginal consumer demand at roughly 51% of global jewellery demand by volume. In 2025, jewellery demand fell 24% while investment demand rose 17%, reaching roughly 40% of total consumption versus a typical quarter.

When sticky demand falls and hot money replaces it, the market looks strong — until the hot money leaves.

Hot Money Is Leaving the Gold Trade

Mon, 23 Mar 2026 00:00:00 GMT

Spot gold plummeting to $4,126 while geopolitical risk is elevated seems counterintuitive. War in the Middle East, oil at $110/barrel — yet gold is down 8–10% in a week. The explanation is liquidity, not fundamentals.

Three forces: The Liquidity ATM effect — when equity markets tank, institutions sell what they can sell. Gold is liquid. It's being treated as a cash source to cover margin calls elsewhere. The hawkish hold trap — energy prices reigniting inflation fears, markets pricing in higher for longer, DXY back toward 100. Speculative exhaustion — the move from $3,000 to over $5,000 was parabolic. Leverage is unwinding.

Factor attribution for 2025: Central bank buying (~20–25% of variance) — inelastic, strategic. 863.3 tonnes, slowing above $4,000. Hot money and ETF flows (~55–60%) — elastic, pro-cyclical. Record 2,175 tonnes, ETFs alone adding 801 tonnes. Macro/geopolitical noise (~15–20%) — reactive.

The marginal price driver shifted from inelastic strategic demand to elastic momentum flows during H2 2025. That explains why the correction is violent: leveraged, stop-loss-governed capital is exiting through a narrow door, while the structural bid sits much lower.

The Gold Bull May Be Temporarily Over

Sun, 22 Mar 2026 00:00:00 GMT

Not necessarily the secular story. But the hot momentum phase in GLD looks finished, or at least paused.

Three things stand out on the chart. The secular bull produced three progressively steeper rising trendlines. The third and steepest has now been broken — typical when the most speculative phase is over. Price shows double-top behavior around $500. Failed breakouts after a parabolic move, in a positive macro environment, may mark exhaustion rather than continuation.

Most importantly, gold failed to make a new all-time high despite a very supportive macro backdrop: geopolitical escalation around Iran and persistent US inflation concerns. When an asset cannot make new highs on bullish news, that is a warning sign.

The base case: a medium-term trend change may be underway. After a steep drop, GLD will probably attempt to recover. But $450 looks like strong resistance. If GLD fails to reclaim that level, the next phase could accelerate to the downside. On the downside, $400 appears plausible in what could become a standard ABC correction.

Gold has been an extremely crowded trade since mid-2025. Unless held for a long-term strategic reason, this is probably not the time to chase. Rallies look more like selling opportunities than invitations to add exposure. The secular bull may still be alive, but the best part of the party looks over.

What Factor Analysis Reveals About Regimes

Sat, 21 Mar 2026 00:00:00 GMT

Standard market commentary identifies what drove returns. Factor analysis identifies how much each driver contributed — and more importantly, whether those drivers are changing.

Applied to Gold (2025): three components explain most of the variance. Central bank buying (~20–25%) providing the structural floor. Hot money and ETF flows (~55–60%) providing the momentum. Macro noise (~15–20%).

The key finding: the β for speculative flows grew roughly 3x larger than the β for central bank buying during H2 2025. The marginal price driver shifted from inelastic strategic demand to elastic momentum flows. That structural shift explains why the correction was violent.

Applied to the Rubin Build-Out: five components proposed. Market beta (~45%), inference pivot (~25%), grid/power (~15%), tariff noise (~10%), idiosyncratic (~5%).

The most important observation is reflexivity: hyperscaler capex announcements feed into NVIDIA's order book, which feeds into NVIDIA's multiple, which feeds into the AI narrative that justifies the capex. That feedback loop is exactly the non-linearity that standard PCA misses.

Correlations between Rubin infrastructure stocks and the broader Nasdaq are regime-dependent — approximately 0.4 in calm markets, approaching 1.0 in a crash. Watching the derivatives of factor participation is where static attribution becomes a dynamic regime detection system.

There Is Always a Bull Market Somewhere

Tue, 24 Feb 2026 00:00:00 GMT

For fifteen years, global capital was hyper-concentrated in U.S. mega-cap tech. That era is ending. VEU has blown through the $64.02 resistance — the 2008 highs that capped every rally for a decade and a half.

Now at $80.95 with RSI embedded at 81.11, the monthly chart shows a parabolic thrust above all three ascending trendlines. Supertrend pristine green at $67.61. Coppock Curve at 47.51 and rising. This is a generational breakout.

VEU holds 3,900 stocks across 46 countries — Financials 21.5%, Industrials 14.5%, hardware-focused Tech 11.5%. It's a bet on a physical, capital-intensive global economy: European banks, Asian foundries, commodity exporters.

The macro fuel: extreme valuation mean reversion, a structural USD bear market, the resurgence of "Old Economy" infrastructure, and Japan/EM reforms. Plus TACO — Tariffs, America-first, Currency weaponisation, Outflows — paradoxically accelerating the rotation.

The 3-ETF Lazy Portfolio (⅓ QQQ, ⅓ VEU, ⅓ ISHG) is up +3.66% YTD on $100K, outperforming QQQ-only by $4,547. Diversification isn't a hedge — in this regime, it IS the return.

King Dollar Cracks: Decade-Long Trend Snaps

Mon, 23 Feb 2026 00:00:00 GMT

The regime change may finally be here. For over a decade, $DXY traded within a massive ascending channel that defined the post-GFC era. That structural support has now given way.

After rejecting 115 in late 2022, the dollar has steadily eroded, finally slicing through the primary trendline support held since 2011. This is not a technical glitch — it's a fundamental shift.

Currently testing the 200-period WMA at 97.07. The Slow Stochastic is deep in the oversold cellar at 12.93 — short-term bounce overdue — but the Coppock Curve has plunged to −13.09, signaling a more secular bear cycle.

Historically, a weakening dollar acts as a tailwind for commodities and emerging markets. With the RSI at 41.98 and Supertrend resistance at 105.57, the path of least resistance remains lower.

If $DXY fails to reclaim the broken channel and loses the 200-MA, the "King Dollar" era may officially be in the history books.

Software Double Bottom? IGV at Support

Fri, 20 Feb 2026 00:00:00 GMT

Software may be ready. IGV completing an ABC correction right into wave (1) support at $82 — the exact level where a double bottom would form. Stochastics washed out across the sector.

The "sell first, ask later" wave hit everything — including names that are structural AI beneficiaries. The market treated all software as disruption victims, but the correction is creating opportunities where AI is additive, not destructive.

Screening for individual stocks showing relative strength against the IGV benchmark. Snowflake stands out: a core data infrastructure play that should benefit from every agentic AI deployment. Agents need data. Agents need governed, structured, queryable data.

Weekly stochastics on SNOW are deeply oversold (20/17) with a clean 5-wave projection off the 2022 lows. Earnings on February 25 — expectations lower, positioning cleaner, AI data narrative stronger.

If IGV holds $82 and SNOW delivers, this is where the software bounce starts.